20th
Research and Creativity
Daniel Lemire’s post on research productivity and the path less traveled strikes a chord with me. He mentions truth-seeking vs. creativity as an example of a traditional vs. non-traditional view of research productivity. I think that creativity isn’t just an interesting “path less traveled”, but rather its essential for real progress to be made.
Describing computer science research as a creative process definitely raises eyebrows, especially when talking to someone outside of the field. Both of my sisters are creative professionals, both with MFAs, one sculptor and one graphic designer. But, I don’t see my work as very different from theirs. We have different tool boxes, work in different mediums and need to produce output under different constraints. My toolbox includes probability theory, theirs includes a needle and thread. My constraints include computing resources, theirs may be the dimensions of a page.
On the surface, one huge difference between my work and theirs is how to quantify success. For the IR research, success is often measured by things like MAP and AUC. There is no analog in the creative professions. But, this is really just one aspect of success, and a particularly shallow one at that. In areas like IR where the user plays a crucial role, a truly successful research project opens the doors for new lines of investigation and makes a positive impact on users of a system. This is exactly the same with creative output — artwork is successful if it inspires other artists and influences the viewer. In some ways, the artists have it much easier in this regard. The physical presences of artwork allows a much more immediate connection with the “user”. For us, there are so many walls around our creative output its hard to know when you’ve done something that will truly have an impact.
